Early Numeracy Skills Mediate the Association Between Relational Language and Mathematical Skills

• In Event: 3-227 - Relational Language and Math: Variation Across Types of Language and Math Activities

Integrative Statement

Children’s knowledge and use of relational language (e.g., more, equal, before, next) correlates with their early numeracy skills in preschool (Purpura & Logan, 2015), and predicts growth in mathematical skills in kindergarten (Toll & Van Luit, 2014). Although relational language can be applied to other mathematical concepts, it may be specifically important for number relation skills, a component of early numeracy skills that involves the knowledge of cardinal (e.g., 5 is more than 4) and ordinal (e.g., 5 comes after 4) number relations. Relational language provides a tool for communicating relations between numbers, and may influence children’s mathematical skills specifically through number relation skills. Here, we examine whether number relation skills mediate the association between relational language and mathematical skills. We test whether the mediation is unique by comparing mediation models substituting knowledge of counting and numerical representations (counting/numbering skills) for number relation skills as the mediator of the association between relational language and mathematical skills.

The participants were 104 kindergarten students (46 boys; mean age: 5.9 years). We assessed children’s skills in relational language, general early mathematics, number relations, and counting/numbering; relational language with the Boehm Test of Basic Concepts – 3rd Edition (Boehm, 2001), and general mathematical skills with the Test of Early Mathematics Ability – 3rd Edition (TEMA; Ginsburg & Baroody, 2003). Number relation skills were measured with tasks that involved ordering and comparing numbers. Counting/numbering skills were measured with tasks that involved verbal counting and numeral identification. We also included covariate measures of general verbal knowledge (the Kaufman Brief Intelligence Test – Verbal Knowledge subtest (KBIT; Kaufman & Kaufman, 2004)), and executive function skills (the Minnesota Executive Function Scale (Carlson & Zelazo, 2014) and Head Toes Knees Shoulders task (Cameron Ponitz et al., 2008)) in analyses. Because children’s skills in number relations, counting/numbering, and executive function were each measured with multiple tasks, the score of each task was scaled to a z distribution and averaged in its respective domain to form a composite score.

The first step of the mediation analysis confirmed that, even with verbal knowledge as a covariate, Boehm performance predicted TEMA scores (Figure 1a). Adding number relation score as a mediator fully mediated that relation (Figure 1b); however, so did counting/numbering (Figure 1c). The pattern of results remained when EF was included as an additional covariate (Figure 1 d - f).

In summary, number relation skills fully mediated the association between relational language and mathematical skills, but so did counting/numbering skills, suggesting that number relation skills and counting/numbering skills may be pathways through which relational language influence mathematical skills. Although the concurrent data precluded causal inferences, reverse mediations with relational language as a mediator between number relation skills (or counting/numbering skills) and mathematical skills were not supported by the data. The inclusion of EF skills as an additional covariate led to smaller path weights between relational language and early numeracy skills but the full mediations remained, suggesting that relational language may share some variance with EF skills but still has independent influences on early numeracy skills.

Authors

• Jenny Yun-Chen Chan, University of Minnesota
  Presenting Author

• Maria Sera, University of Minnesota
  Non-Presenting Author

• Michele Mazzocco, University of Minnesota
  Non-Presenting Author